Forum:We've done a Wiki Merger, and this is how

Forums: Index > Village Dump > We've done a Wiki Merger, and this is how

Note: This topic has been unedited for 1864 days. It is considered archived - the discussion is over.

Binky The WonderSkull and I, with help from others, merged two wikis, Society for Creative Anachronism Wiki and SCA Wiki. We did it by the export/import categories method. It worked, but it was slow and tedious.

I co-admin another wiki where we want to back the whole thing up. But it's significantly larger than the wikis we merged. I've asked for suggestions, and have gotten the one I've done from a helpful user (I would credit the person but don't know their Uncyclopedia name), and suggestions from someone who knows a lot more about computer tech stuff than I do. The only one of those suggestions I'm actually capable of doing is hiring a tech to do the backup.

Anybody know an easy way for a non-computer tech to backup a wiki without paying someone to do it? (I know enough to have successfully run a school computer lab many years ago, but that's about it).  King of the Internet Alden Loveshade??? (royal court)  19:05, March 18, 2019 (UTC)~

The words "easy" and "backup" and "wiki" don't belong together in one sentence (assuming, of course, that you plan on restoring it again from those backups some time in the future).

An xml dump, which I believe you can get from inside the wiki, doesn't include assorted random items like logs, which you will want if you plan on doing a disaster-recovery restore.

As I understand it, to get the whole thing you need an sql database dump, which you can't get without root access to the host system. And even then you need to dump the images separately, because they're not part of the database. And then I'm still not sure that's everything. (My involvement with wiki hosting has mostly been dealing with the underlying system, not the wiki itself, so someone else who knows more may have more encouraging words for you -- but I doubt it.)

The grief you have go to through to import a full wiki and get everything (rather than just most of it) is really rather remarkable, given how many years mediawiki has been around working on scripts to do stuff like this. The one really, really simple approach that comes to mind is to host it on a virtual machine and then snapshot the virtual machine -- but aside from the issue of bulk (that's sure not compact!), you can't do that at all without console access to the VM, which you probably don't have, unless you own the physical server. Snarglefoop (talk) 21:47, March 18, 2019 (UTC)

If you have root access to the host system you can just snapshot /var/www and /var/lib, and that gets you pretty much everything (except maybe those pesky images). You probably want to lock the database (or stop Apache) while you're doing that, of course (which may not make the users happy). And you don't want to try to import that "snapshot" into any system that isn't essentially identical to the one you got it from. (It's a useful dump to have if, say, your main server blows its power supply.) Snarglefoop (talk) 21:57, March 18, 2019 (UTC)

And if your main server DOES have its power supply let go, you probably either want the server to be off-site, or you want it to happen during the summer, because you're going to need to open the windows afterwards. It makes a truly ferocious stench. (Just went through that month or so back here. Ugh.) Snarglefoop (talk) 22:00, March 18, 2019 (UTC)

I think this task should be split into as many administrators as possible who can handle those tools. Rhubella Marie, the rat sockpreppie preppiedits 01:20, March 19, 2019 (UTC)

That's always good, but where do you find people willing to do it? ❦ Llwy-ar-lawr • talk • contribs • 07:29 19 March 2019

You want to back up a Wikia wiki, which you can't do from the server end. It can be done, but it depends on what you mean by "easy" and "the whole thing".

First you'll want to get a dump of the content from Wikia. This is easy. These can be downloaded from the statistics page. I don't see any download links, just "unknown", so I think you'll have to request one. You're an admin, so there should be a button for you. Once you do that and the link appears, you can just click it and download the dump. You probably want the "current pages and history" one.

That will only get you the pages and histories, not the images or anything else. If you want any of that, it gets harder.

You can get the images (current versions only) with WikiTeam. This is fairly straightforward but requires some sort of command line, digging through instructions to figure out how to use it, and probably installing the necessary dependencies. I had to do a lot of figuring out here. I'm not really a technical person either. I had to follow the instructions here because it complained that the "kitchen module" wasn't installed, then follow the instructions here when it gave the error "Badly placed ()'s". Then I was able to run it. The tutorial is here; you'll type in something like python dumpgenerator.py --api=http://trynser.fandom.com/api.php --images.

If you want more than that, the only way I know is to use mw:Manual:Grabbers and download the stuff onto a fresh MediaWiki installation. Installing MediaWiki is the easy part (the page shows how to do it in a single command). The scripts are fairly easy to run except when they run into trouble and stop dead, which is often, and I still haven't totally figured them out yet. Also, you will not get the user list, and supposedly you won't get abuse filter histories or user rights either.

If you already knew all that, then never mind. Hope something in there was remotely helpful. ❦ Llwy-ar-lawr • talk • contribs • 07:29 19 March 2019

I appreciate everyone's advice. I knew nothing better than the export/import option, Some of this may take me some work to figure out and I may ask for help along the way, but I believe you've given me a good starting point. Thank you very much!  King of the Internet Alden Loveshade??? (royal court)  04:13, March 20, 2019 (UTC)